Why pay attention to ROE? attention, actually, capital, company, people

cyber21

OK guys, I'm embarrassed to be asking this, but I figured some of you ER people can explain how to look at ROE?

I have two issues that make me uncomfortable paying much attention to ROE.

1) ROE uses equity book values rather than market values, so you may think that a company is using equity capital effectively, but the earnings yield based on market prices is really rather blah.

2) ROE is strongly affected by financial leverage, so ROE might look great, but there's actually substantial risk attached to it.



I understand ROA and ROIC as measures of company operating efficiency, and I like those metrics, but what do you get out of ROE that you wouldn't get by taking a look at earnings yields based on market prices?

dmrktrading

> what do people look at for balance sheet strength

Three things: credit rating, credit rating, and credit rating.

Any metric you can dream up, the agency analysts have already considered, tried, and tested for relevance to that particular sector. Don't spend time reinventing this wheel.

torontoanalyst

is there a screen for return on tangible book equity? That would be a more valid screen due to goodwill, intangible asset considerations that should be excluded.

therecruit

It's interesting. The macro stuff comes so naturally to me. Company level analysis is something where I'm pretty confident I know the tools and am smart enough to do it, but it feels a bit like learning a new dance step, when you're asking yourself "am I sure I'm doing it right? Am I forgetting something important that's going to mess me up? Is this partner going to totally barf up all over me because I didn't realize she'd had way too much to drink."

But I'm pretty sure we're getting to a particular moment soon where it going to be especially key to separate the wheat from the chaff at the company level. Macro will still be important, but if you want that extra juice, time to look at the company level.

cfa10yrplan

Thanks, Carson

Yeah, I like ROIC.

BTW, I'm not saying I disbelieve the value of ROE ; I'm just trying to get a better intuitive feel for what it adds to the analysis vs things like ROIC and earnings yield.


And, while we're at it, what do people look at for balance sheet strength? I'd be tempted to look at 1) D/E ratios, 2) quick ratio, and 3) something like volatility of earnings divided by book equity (maybe that is pretty similar to the volatility of ROE).

canadiananalyst

Palantir Wrote:
-------------------------------------------------------
> I think it's especially useful in screening if you
> look at RoA side by side with RoE. If a firm has
> historically generated RoA and RoE of say between
> 10-15 very consistently, I think you can make
> confident projections of earnings down the line.
>
>
> Another thing to keep in mind while screening is
> to also be looking at how much cash they have per
> share as well as debt. Cash brings down the
> displayed RoA/E, but it could really be masking a
> much higher one. I like to look for firms that
> have a lot of cash+equivalents on hand, but with a
> relatively meager RoA.

I agree with this. I always screen for both ROE and some measure of balance sheet strength. Gearing is one factor in computing ROE (ala DuPont analysis which you mentioned) so you want to look for companies with high ROE but low levels of debt. One way to do that is to screen for high ROE and ROIC (or ROA if you like) simultaneously. That eliminates your second reservation.

ROE is a good indicator of the efficiency of a company I think. Firms with high ROE and high ROIC (say >15% over the medium term) will tend to have some sort of competitive advantage which should imply that earnings and cashflow are of a high quality and are likely to be sustainable into the future. That's not always the case of course, but ceteris paribus it is a good indicator.

And if that doesn't convince you that ROE is a good metric to look at, remember that it is one of the first things Warren Buffet looks at and he is a pretty sharp investor.

dmrktrading

> Sometimes ROE fails to predict future growth rates because the marginal ROE on new projects is quite a bit lower.

There are other issues to consider. E.g. book value of assets might be quite different from market price of them should you wish to invest in more. A more common issue is that most firms don't automatically grow nominal amount of debt as they reinvest, meaning their leverage is declining and ROE will suffer.

I think ROA is always inferior to ROIC, given that cash on balance sheets is usually mostly surplus. (However either of these allows easier cross-firm comparisons due to the leverage skewing in ROE that you point out.)

My gut feeling is that ROE (or rote or roce) is used more in financial industries where book and mkt values of assets are much closer, and in regulated subsectors you'll find leverage levels very similar across competitors.

wizofoz

I think it's especially useful in screening if you look at RoA side by side with RoE. If a firm has historically generated RoA and RoE of say between 10-15 very consistently, I think you can make confident projections of earnings down the line.


Another thing to keep in mind while screening is to also be looking at how much cash they have per share as well as debt. Cash brings down the displayed RoA/E, but it could really be masking a much higher one. I like to look for firms that have a lot of cash+equivalents on hand, but with a relatively meager RoA.

wizofoz

Good stuff. I remember always looking at ROE on CFA L1 problems to get at the growth rate for reinvested stuff.

I also remember that the Dupont model was pretty useful for figuring out whether ROE was coming from turnover or from profit margins or from financial leverage.


Palantir's point about persistence of earnings is good too. I hadn't thought of that. Valuations can fluctuate a lot without necessarily being related to company operations, so that would affect earnings yields but not ROE.


I suppose a big thing is that book value might be (for some types of companies) a more solid basis for projecting future earnings than market value of equity (though I'd still be inclined to use ROA for that). I guess that's a kind of corollary to both 99 c sloop and Palantir's comments, but I never quite got that when I was studying for CFA.


Anything else? I know some of you guys pay attention to ROE fairly regularly.

(I'm thinking about this as I try to work through good stock screens for the current environment, and know that people often use ROE as a criterion)

zbird2134

I think it's really about trying to figure out whether earnings are persistent. I don't think earnings yield would give you that info.